Added commandline tool for runing prediction

Author: bafgreat

.gitignore

@@ -246,4 +246,7 @@ __pycache__/
 .pytest_cache/
 .ipynb_checkpoints/
 examples/create_metal_salts.py
-examples/test.xyz
+examples/test.xyz
+examples/iupacname_smiles.msgpack
+fairmofsyncondition/call_model/mof_prediction.txt
+examples/combine_all_names2.py

README.md

@@ -53,6 +53,22 @@ pip install fairmofsyncondition
 
 ## Useful tool
 
+`fairmofsyncondition_syncon`
+is a command-line tool for predicting synthetic conditions of Metal–Organic Frameworks (MOFs) directly from CIF files.
+It extracts **organic ligands**, **space group information**, and computes the **top-5 predicted metal salts**.
+
+Quickly run command on any cif file
+
+```bash
+fairmofsyncondition_syncon .my_mof.cif
+```
+
+Or run and provide and outfile
+
+```bash
+fairmofsyncondition_syncon my_mof.cif -o my_mof_report.txt
+```
+
 `iupac2cheminfor`
 one of the most useful tool is to directly extract cheminonformatic identifiers such
 as inchikey and smile strings directly from iupac names or common names. This can be

docs/source/index.rst

@@ -45,6 +45,21 @@ To install `fairmofsyncondition` from PYPI, simply execute the following command
 
     pip install fairmofsyncondition
 
+
+Predict Synthesis Conditions
+============================
+
+The quickest way to predict the synthesis condition from any crystal structure is by
+running the below commandline argument.
+
+.. code-block:: bash
+    fairmofsyncondition_syncon my_mof.cif
+
+.. code-block:: bash
+    fairmofsyncondition_syncon my_mof.cif -o my_mof_report.txt
+
+
+
 License
 =======
 

fairmofsyncondition/call_model/run_model.py

@@ -14,6 +14,7 @@
 from fairmofsyncondition.read_write import filetyper, coords_library
 from fairmofsyncondition.call_model.utils_model import get_models, ensemble_predictions
 
+smile_to_iupac = filetyper.smile_names_iupac()
 
 convert_struct = {'cubic': 0,
                   'hexagonal': 1,
@@ -26,36 +27,37 @@
 
 device = "cpu"
 
-def get_ligand_iupacname(ligand_inchi):
+def get_ligand_iupacname(ligand_inchi, smiles):
     name = load_iupac_names().get(ligand_inchi, None)
-    print("tesst", name)
 
     if name is None:
-        pubchem = cheminfo2iupac.pubchem_to_inchikey(ligand_inchi, name='inchikey')
+        name = smile_to_iupac.get(smiles, None)
+    if name is None:
+        pubchem = cheminfo2iupac.pubchem_to_inchikey(smiles, name='smiles')
         if pubchem is None:
-            name = ligand_inchi
+            name = None
         else:
-            pubchem.get('iupac_name', ligand_inchi)
+            name = pubchem.get('iupac_name', None)
     return name
 
 class Synconmodel(object):
     def __init__(self,
                  ase_atom,
-                 outfile="mof_prediction.txt",
-                 module="best_model.pth",
+                 outfile="mof_prediction.txt"
                  ):
         """
         A class to run the metal salt prediction GNN model and for
         extracting organic ligands and predicting metal salts.
         """
         if isinstance(ase_atom, Atoms):
             self.ase_atom = ase_atom
+            self.basename = 'MOF Data'
         elif isinstance(ase_atom, str) and os.path.isfile(ase_atom):
             self.ase_atom = read(ase_atom)
+            self.basename = os.path.basename(ase_atom).split('.')[0]
         else:
             print("sorry input type is not recorgnised")
         self.outfile = outfile
-        self.gnn_model = module
         self.structure = MOFstructure(self.ase_atom)
         self.pymat = AseAtomsAdaptor.get_structure(self.ase_atom)
         self.torch_data = coords_library.ase_to_pytorch_geometric(self.ase_atom)
@@ -103,11 +105,12 @@ def get_coordination_and_oms(self):
         return emb, oms, metals
 
     def get_organic_ligands(self):
+        ligands_data = {}
         _, ligands = self.structure.get_ligands()
         inchikeys = [ligand.info.get('inchikey') for ligand in ligands]
-        ligands_names = [get_ligand_iupacname(i) for i in inchikeys]
-        print(ligands_names)
-
+        smiles = [ligand.info.get('smi') for ligand in ligands]
+        ligands_names = [get_ligand_iupacname(i, j) for i, j in  zip(inchikeys, smiles)]
+        return inchikeys, smiles, ligands_names
 
     def get_space_group(self):
         "Get space group embedding"
@@ -118,13 +121,13 @@ def get_space_group(self):
         emb_sg[space_group_number] = 1
         get_crystal_system = sga.get_crystal_system()
         emb_cs[convert_struct[get_crystal_system]] = 1
-        return emb_sg, emb_cs
+        return emb_sg, emb_cs, space_group_number, get_crystal_system
 
     def complete_torch_data(self):
         '''
         Get torch data
         '''
-        emb_sg, emb_cs = self.get_space_group()
+        emb_sg, emb_cs, _, _ = self.get_space_group()
         cn_emb, oms, metals = self.get_coordination_and_oms()
         atom_conc = self.get_species_conc()
         self.torch_data.atomic_one_hot = atom_conc
@@ -134,27 +137,124 @@ def complete_torch_data(self):
         self.torch_data.oms = torch.tensor([[oms]], dtype=torch.float)
         return self.torch_data.to(device), metals
 
-
     def predict_condition(self):
         '''
         predict conditions
         '''
         torch_data, metals = self.complete_torch_data()
         torch_data = torch_data.to(device)
-        torch_data.cordinates = torch_data.cordinates[0] # small cleaning
-        models = get_models(torch_data, device=device) # load models (5 seeds)
+        torch_data.cordinates = torch_data.cordinates[0]
+        models = get_models(torch_data, device=device)
         models = models[0:1]
 
         category_names = filetyper.category_names()["metal_salts"]
-        # the function to get ensemble predictions, i.e. the average probability for each class over the 5 models
         pred_list = ensemble_predictions(models, torch_data, category_names, device=device)
-
-        for name, prob in pred_list[:10]:
-            print(f"{name}: {prob:.3f}")
+        # for name, prob in pred_list[:10]:
+        #     print(f"{name}: {prob:.3f}")
 
         #print(pred_list)
+        return pred_list
+
+    def compile_data(self):
+        data = []
+        inchikeys, smiles, ligands_names = self.get_organic_ligands()
+        _, _, space_group_number, get_crystal_system = self.get_space_group()
+        pred_list = self.predict_condition()[:5]
+
+        data.append("\n")
+        data.append(f"Predicted Synthetic Data Report\n")
+        data.append(f"For: {self.basename}\n")
+        data.append("=" * 80 + "\n")
+        data.append(f"{'Space group number:':25} {space_group_number}\n")
+        data.append(f"{'Crystal system:':25} {get_crystal_system}\n\n")
+
+
+        data.append("Organic Ligands\n")
+        data.append("-" * 80 + "\n")
+        data.append(f"{'InChIKey':<28} {'SMILES':<30} {'IUPAC Name':<20}\n")
+        data.append("-" * 80 + "\n")
+        for inchi, smi, iupac in zip(inchikeys, smiles, ligands_names):
+            inchi_str = str(inchi) if inchi is not None else "N/A"
+            smi_str = str(smi) if smi is not None else "N/A"
+            iupac_str = str(iupac) if iupac is not None else "N/A"
+            data.append(f"{inchi_str:<28} {smi_str:<30} {iupac_str:<20}\n")
+        data.append("\n")
+
+
+        data.append("Top 5 Predicted Metal Salts\n")
+        data.append("-" * 80 + "\n")
+        data.append(f"{'Metal Salt':<40} {'Probability':>15}\n")
+        data.append("-" * 80 + "\n")
+        for metal_salt, prob in pred_list:
+            salt_str = str(metal_salt) if metal_salt is not None else "N/A"
+            prob_str = f"{prob:.4f}" if prob is not None else "N/A"
+            data.append(f"{salt_str:<40} {prob_str:>15}\n")
+        data.append("=" * 80 + "\n")
+
+        data.append("\n")
+        data.append("Report generated by fairmofsyncondition\n")
+        data.append("Authors: Dinga Wonanke & Antonio Longa\n")
+        data.append("=" * 80 + "\n")
+
+
+        filetyper.put_contents(self.outfile, data)
+
+
+def print_helpful_information():
+    '''
+    Prints helpful information about using the fairmofsyncondition script.
+    '''
+    help_text = """
+    Usage: fairmofsyncondition.py [CIF_FILE] [OPTIONS]
+
+    This script predicts synthetic conditions for MOFs given a CIF file
+    or any ase readable fileformat.,
+    extracts organic ligand information, space group, and top predicted
+    metal salts, and writes a formatted report to a text file.
+
+    Positional Arguments:
+      CIF_FILE              Path to the input CIF file.
+
+    Optional Arguments:
+      -o, --output FILE     The path to the output report file (default: prediction_report.txt).
+
+    Examples:
+      fairmofsyncondition_syncon my_mof.cif -o my_mof_report.txt
+      fairmofsyncondition_syncon ../data/sample.cif
+    """
+    print(help_text)
+
+def main():
+    parser = argparse.ArgumentParser(
+        description="Predict synthetic conditions for MOFs from CIF files."
+    )
+    parser.add_argument(
+        'cif_file',
+        type=str,
+        nargs='?',
+        default=None,
+        help='Path to the CIF file to analyze.'
+    )
+    parser.add_argument(
+        '-o', '--output',
+        type=str,
+        default='mof_prediction_report.txt',
+        help='Path to the output report file (default: prediction_report.txt).'
+    )
+
+    args = parser.parse_args()
+
+    if args.cif_file is None:
+        print_helpful_information()
+        sys.exit(1)
 
+    try:
+        ml_data = Synconmodel(args.cif_file)
+        ml_data.outfile = args.output
+        ml_data.compile_data()
+    except Exception as e:
+        print(f"Error: {e}")
+        sys.exit(1)
 
-ml_data = Synconmodel('../../tests/test_data/EDUSIF.cif')
-# print(ml_data.predict_condition())
-print(ml_data.get_organic_ligands())
+if __name__ == '__main__':
+    main()

fairmofsyncondition/db/iupacname_smiles.msgpack

@@ -40,7 +40,12 @@ def pubchem_to_inchikey(identifier, name='smiles'):
                       otherwise, None.
     '''
     properties = {}
+
     pubchem = pcp.get_compounds(identifier, name)
+    # print('inchikey', pubchem[0].inchikey)
+    # print('name', pubchem[0].synonyms)
+    # print('iupac', pubchem[0].iupac_name)
+    # print('isomeric_smile', pubchem[0].isomeric_smiles)
     if len(pubchem) > 0:
         all_prop = pubchem[0]
         properties['inchikey'] = all_prop.inchikey if all_prop.inchikey else None

fairmofsyncondition/read_write/cheminfo2iupac.py

@@ -560,3 +560,9 @@ def category_names():
     "load category"
     msgpack_path = files("fairmofsyncondition").joinpath("db/category_names.msgpack")
     return load_data(msgpack_path)
+
+
+def smile_names_iupac():
+    "load iupac"
+    msgpack_path = files("fairmofsyncondition").joinpath("db/iupacname_smiles.msgpack")
+    return load_data(msgpack_path)

fairmofsyncondition/read_write/filetyper.py

@@ -42,6 +42,7 @@ find_bde_parameters = "fairmofsyncondition.model.hyper_optimiser:main"
 iupac2cheminfor = "fairmofsyncondition.read_write.iupacname2cheminfo:main"
 cheminfor2iupac = "fairmofsyncondition.read_write.cheminfo2iupac:main"
 struct2iupac = "fairmofsyncondition.read_write.struct2iupac:main"
+fairmofsyncondition_syncon = "fairmofsyncondition.call_model.run_model:main"
 
 
 [tool.poetry.group.dev.dependencies]